Changing demographic rates entail eventual changes in the numbers and types of living kin that elderly persons will have and thus in their potential kin-based resources. Computer microsimulation is a tool for projection these changes in kinship resources based on assumptions about the future course of demographic rates. This project will begin with an external validity study, believed to be the first ever conducted for kinship projection, comparing earlier projections of 1980s kinship against empirical estimates from the National Survey of Families and Households (NSFH). Next the effect of employing different modelling techniques will be assessed by comparing two sets of Chinese kinship forecasts, one derived by computer microsimulation and the other by the alternative technique of family status life tables. The microsimulation projection model and U.S. historical rate sets will be revised on the basis of these tests. NSFH information of kin counts will be extrapolated to the over-80 population using simulation where sample sizes are insufficient for stable estimates. New kinship projections for the U.S. will then be prepared down to 2050, separately for blacks and for whites and others, using a range of alternative forecasts of future demographic rates. The NSFH makes it possible to assess the degree to which kin of different types within kinship congfigurations of different types actually do now provide support of various kinds. This will allow the construction and projection of kinship-based resource indices for the United States which go beyond kin counts. Special emphasis will be placed on assessment of kinship networks for the "oldest old".